Exploring habitat, community, and meta-community assembly in temperate wooded headwater streams after destructive flooding

Background/Question/Methods

The major objective of this study is long-term quantification of physical and biotic environments within and among 14 closely adjacent 1^st – 3^rd-order wooded streams in western New York State USA, following their extensive destruction by a flash flood in 2009. Sites represent a sub-set of streams in which watershed and in-stream environmental characteristics and macroinvertebrate biota (to genus) were quantified in 2006, and published in the peer-reviewed literature. Many streams originally possessed heterogeneous in-channel structure and diverse biota, but after the flood they were largely scoured of substrata and woody debris, with some channels left nearly barren. Biotic and environmental assessments from 2010 and onward will mirror those from 2006 (with modifications as warranted), with biota collected from riffle/cobble segments by Surber net, and environmental/habitat variables quantified by maps and aerial photos, ground-truthing, and a widely-used Qualitative Habitat Evaluation Index (QHEI) that assesses in-stream and riparian characteristics. In addition to yearly descriptive data, this study seeks to investigate spatio-temporal trends in meta-community assembly in terms of environmental (likely niche-based) and/or spatial structuring (potentially equivalence-based neutral theory or homogeneous patch dynamics). 

Results/Conclusions

Changes in both environmental and biotic structuring among steams have occurred between pre- and post-flood samples. In-stream environmental variables (e.g. substrate type and diversity, in-stream cover) converged, as did QHEI scores that formerly discriminated stream orders clearly. This suggests the development of more homogeneous post-flood “patches”. A clear discrimination in 2006 among stream orders in Principle Components Analysis of environmental variables was lost by 2010. A similar pre-flood clustering of 2^nd and especially 3^rd order streams based on ordination of  the biota by Non-metric Multi-dimensional Scaling was also lost as of 2010. In 2006, partial Mantel correlations indicated a statistically significant association between environmental and biotic Euclidean distance matrices, but no such correlation with the spatial matrix. Conversely, by 2010, a significant spatial/biotic correlation had emerged. Pre-flood trends from 2006 suggested a heterogeneity/niche-based meta-community assembly, hypothesized to represent species sorting, although with the possibility of a more dispersal-associated mass effects model. By 2010, however, spatially-structured meta-community assembly instead appeared more likely. A demonstrated loss of among-stream environmental distinctness may favor homogeneous patch dynamics, although if biotic composition has also become constrained, and perhaps more disturbance tolerant, equivalence-based neutral theory could also operate.